Five experiments are proposed to investigate the hemispheric nature of long-term semantic memory (LTSM). The experiments are designed for three subject populations: Stroke patients with unilateral focal lesions, (complete) commissurotomy patients, and normal subjects. The stimuli are pictorial and the responses too are non-verbal. The experiments measure the semantics of complex pictorial scenes, of pictures of individual natural objects, of artificial pictorial concepts, and of logically defined concepts, as follows: (1) Test memory for scenes when the scenes have normal organization with respect to everyday life i.e., familiar schemata (organized), lack organization, i.e., randomly juxtaposed objects (unorganized), or violate coherence by portraying parts of two conflicting schemata in the same picture (incongruous); (2) determine "mental distances" among concepts of natural superordinate categories with typicality levels of common objects serving as the guideposts; (3) study formation of "prototypical" concepts with the use of artificial (random-dot patterns) categories; (4) investigate the validity of the hypothesis that the left hemisphere uses a logic-bound classificatory system and that the right uses an experience-bound system for all concepts alike by the use of logically- defined categories, e.g., odd-number. Characterizing the mechanisms of functional asymmetries in the human brain is important for both basic scientific research and medical practice. Unilateral stroke, tumor, or missle wounds can cause devastating language, cognitive, and memory disorders and treatment for such impairments can improve only with increased understanding of hemispheric specialization. The experiments proposed here should help gain direct insight into hemispheric mechanisms of a fundamental central processing module, namely, the conceptual system, LTSM, which plays a central role in modern cognitive scientific research.